2 citations · 3 across the 26 of their papers we have counts for
26 papers
GeoBalance: Geometry-Aware Monitoring and Reconstruction with Asymmetric Optimization for Balanced Multimodal Learning
Zechang Xiong, Da Li, Rong Yin +7
Multimodal classifiers can converge to modality-dominant solutions in which one modality dominates the joint prediction, suppressing the learning of others. Existing balancing meth…
Can LVLMs Uncover the Truth Behind Visual Illusions? An Analysis of Perceptual and Reasoning Capabilities
Liangjie Zhao, Jiaqing Lyu, Kexin Tang +5
Large Vision Language Models have integrated reasoning capabilities, elevating cognitive performance to new levels. However, existing evaluations either focus solely on perception…
Dense Supervision, Sparse Updates: On the Sparsity and Geometry of On-Policy Distillation
Guo Yu, Wenlin Liu, Yulan Hu +3
On-policy distillation (OPD) has recently become a prominent post-training recipe by combining two desirable ingredients: on-policy student-generated trajectories and dense token-l…
GroupTravelBench: Benchmarking LLM Agents on Multi-Person Travel Planning
Xiang Cheng, Yulan Hu, Lulu Zheng +4
Travel planning in the real world is overwhelmingly a \textit{group} activity, yet existing LLM travel-planning benchmarks reduce it to a single user, where the field is approachin…
Multi-Stakeholder LLM Alignment: Decomposing Estimation from Aggregation
Lulu Zheng, Wenjin Yang, Xiangwen Zhang +4
Multi-stakeholder tasks require one output to satisfy users with conflicting preferences. Holistic LLM judges conflate utility estimation and utility aggregation, yielding unstable…
STEPS: Selective On-Policy Self-Distillation for Reasoning
Jiaxuan Wang, Xuan Ouyang, Zhiyu Chen +4
On-policy self-distillation (OPSD) uses a model as its own teacher under privileged context, providing token-level supervision on the model's own reasoning trajectories. However, p…
Aligning Agents via Planning: A Benchmark for Trajectory-Level Reward Modeling
Jiaxuan Wang, Yulan Hu, Wenjin Yang +3
In classical Reinforcement Learning from Human Feedback (RLHF), Reward Models (RMs) serve as the fundamental signal provider for model alignment. As Large Language Models evolve in…
Beyond Forced Modality Balance: Intrinsic Information Budgets for Multimodal Learning
Zechang Xiong, Da Li, Kexin Tang +3
Multimodal models often converge to a dominant-modality solution, in which a stronger, faster-converging modality overshadows weaker ones. This modality imbalance causes suboptimal…
MapTab: A Diagnostic Benchmark for Long-Horizon Multi-Criteria Multimodal Reasoning on Heterogeneous Topological Graphs
Ziqiao Shang, Ling-Yue Ge, Zian Xu +10
Systematically evaluating Multimodal Large Language Models (MLLMs) is essential for advancing Artificial General Intelligence (AGI). Yet existing benchmarks remain inadequate for r…
Learn More with Less: Uncertainty Consistency Guided Query Selection for RLVR
Hao Yi, Yulan Hu, Xin Li +3
Large Language Models (LLMs) have recently improved mathematical reasoning through Reinforcement Learning with Verifiable Reward (RLVR). However, existing RLVR algorithms require l…
AMAP Agentic Planning Technical Report
AMAP AI Agent Team, Yulan Hu, Xiangwen Zhang +22
We present STAgent, an agentic large language model tailored for spatio-temporal understanding, designed to solve complex tasks such as constrained point-of-interest discovery and…
No More Stale Feedback: Co-Evolving Critics for Open-World Agent Learning
Zhicong Li, Lingjie Jiang, Yulan Hu +7
Critique-guided reinforcement learning (RL) has emerged as a powerful paradigm for training LLM agents by augmenting sparse outcome rewards with natural-language feedback. However,…
Beyond Itinerary Planning-A Real-World Benchmark for Multi-Turn and Tool-Using Travel Tasks
Xiang Cheng, Yulan Hu, Xiangwen Zhang +5
Travel planning is a natural real-world task to test large language models' (LLMs) planning and tool-use abilities. Although prior work has studied LLM performance on travel planni…
Towards Reward Fairness in RLHF: From a Resource Allocation Perspective
Sheng Ouyang, Yulan Hu, Ge Chen +3
Rewards serve as proxies for human preferences and play a crucial role in Reinforcement Learning from Human Feedback (RLHF). However, if these rewards are inherently imperfect, exh…
SPPD: Self-training with Process Preference Learning Using Dynamic Value Margin
Hao Yi, Qingyang Li, Yulan Hu +3
Recently, enhancing the numerical and logical reasoning capability of Large Language Models (LLMs) has emerged as a research hotspot. Existing methods face several limitations: inf…
Coarse-to-Fine Process Reward Modeling for Mathematical Reasoning
Yulan Hu, Sheng Ouyang, Jinman Zhao +1
The Process Reward Model (PRM) plays a crucial role in mathematical reasoning tasks, requiring high-quality supervised process data. However, we observe that reasoning steps genera…
Video-Text Dataset Construction from Multi-AI Feedback: Promoting Weak-to-Strong Preference Learning for Video Large Language Models
Hao Yi, Qingyang Li, Yulan Hu +3
High-quality video-text preference data is crucial for Multimodal Large Language Models (MLLMs) alignment. However, existing preference data is very scarce. Obtaining VQA preferenc…
GUNDAM: Aligning Large Language Models with Graph Understanding
Sheng Ouyang, Yulan Hu, Ge Chen +1
Large Language Models (LLMs) have achieved impressive results in processing text data, which has sparked interest in applying these models beyond textual data, such as graphs. In t…
TSO: Self-Training with Scaled Preference Optimization
Kaihui Chen, Hao Yi, Qingyang Li +4
Enhancing the conformity of large language models (LLMs) to human preferences remains an ongoing research challenge. Recently, offline approaches such as Direct Preference Optimiza…
Preserving Node Distinctness in Graph Autoencoders via Similarity Distillation
Ge Chen, Yulan Hu, Sheng Ouyang +2
Graph autoencoders (GAEs), as a kind of generative self-supervised learning approach, have shown great potential in recent years. GAEs typically rely on distance-based criteria, su…
Towards Comprehensive Preference Data Collection for Reward Modeling
Yulan Hu, Qingyang Li, Sheng Ouyang +6
Reinforcement Learning from Human Feedback (RLHF) facilitates the alignment of large language models (LLMs) with human preferences, thereby enhancing the quality of responses gener…
Exploring Task Unification in Graph Representation Learning via Generative Approach
Yulan Hu, Sheng Ouyang, Zhirui Yang +4
Graphs are ubiquitous in real-world scenarios and encompass a diverse range of tasks, from node-, edge-, and graph-level tasks to transfer learning. However, designing specific tas…
VIGraph: Generative Self-supervised Learning for Class-Imbalanced Node Classification
Yulan Hu, Sheng Ouyang, Zhirui Yang +1
Class imbalance in graph data presents significant challenges for node classification. While existing methods, such as SMOTE-based approaches, partially mitigate this issue, they s…
Graph Ranking Contrastive Learning: A Extremely Simple yet Efficient Method
Yulan Hu, Sheng Ouyang, Jingyu Liu +6
Graph contrastive learning (GCL) has emerged as a representative graph self-supervised method, achieving significant success. The currently prevalent optimization objective for GCL…